Topics › System
Biostatistics & Epidemiology/Population Health — Step 1 practice questions
190 questions in this area, with an explanation for every answer choice.
Three samples below, free and without an account.
How this shows up on Step 1
Reliably the most learnable section on the exam. Study design, the 2x2 table, and the difference between sensitivity and predictive value account for most questions, and none of it requires clinical knowledge. Students who leave it until the end lose easy marks.
Sample questions
Sample question 1
A study with a small sample size reports a very wide 95% confidence interval for its effect estimate. Which of the following best explains this finding?
Explanations
A. Sample size has no relationship to confidence interval width
Sample size has a well-established, direct relationship to confidence interval width — larger samples generally produce narrower, more precise intervals.
B. Wide confidence intervals always indicate a fraudulent study
A wide confidence interval reflects statistical uncertainty due to sample size/variability, not necessarily fraud.
C. Confidence interval width is determined solely by the significance level and nothing else
While significance level (e.g., 95% vs 99%) affects interval width, sample size and variability are also major determining factors, not just significance level alone.
D. A wide confidence interval means the study found no effect at all, definitively
A wide confidence interval reflects imprecision, not necessarily "no effect" — the interval could still exclude the null value despite being wide, or could span both meaningful and null effects, requiring careful interpretation.
E. Smaller sample sizes generally produce less precise (wider) confidence intervals due to greater sampling variability · correct
Smaller sample sizes are associated with greater sampling variability, producing wider (less precise) confidence intervals around the point estimate, reflecting greater uncertainty about the true population value.
Sample question 2
In a survival analysis, a participant who is alive and disease-free when the study ends, and another who moves away and is lost to follow-up at 3 years, are both marked with tick marks on the Kaplan-Meier curve. Which of the following best describes the assumption underlying this treatment of the data?
Explanations
A. Censored participants are assumed to have experienced the outcome at the time of censoring
Assuming the event occurred would overestimate the hazard.
B. Censoring requires that all participants enter the study on the same calendar date
Staggered entry is precisely what survival methods accommodate.
C. Censoring is assumed to be noninformative — that censored participants have the same subsequent risk as those who remain under observation · correct
Kaplan-Meier analysis accommodates participants with incomplete follow-up by censoring them at their last known event-free time; they contribute to the risk set until then. The key assumption is that censoring is noninformative — unrelated to prognosis. If sicker patients drop out preferentially, survival estimates become biased.
D. Censored participants are excluded from the analysis entirely from time zero
Censored participants contribute information until their censoring time; they are not discarded from the outset.
E. Censored participants are assumed never to experience the outcome
Assuming immunity would underestimate it.
Sample question 3
Researchers enroll a cohort of smokers and non-smokers and follow them forward for 20 years to determine the incidence of lung cancer in each group. Which measure of association can this study calculate that a case-control study cannot?
Explanations
A. Correlation coefficient
A correlation coefficient measures a linear relationship between two continuous variables, not the standard risk measure for this type of exposure-outcome study.
B. Neither can calculate any risk measure
Cohort studies absolutely can and do calculate risk measures, including relative risk.
C. Odds ratio
Odds ratios can be calculated by BOTH cohort and case-control studies; this is not the distinguishing measure.
D. Relative risk · correct
A cohort study, by following defined groups forward and observing actual disease incidence, can calculate relative risk directly — something a case-control study (which starts with a fixed number of cases/controls) cannot do.
E. Prevalence ratio only
While prevalence can be measured in some cross-sectional contexts, incidence-based relative risk is the key distinguishing measure for a cohort study here.
Practice the rest
Build a quiz from this topic, or sit a full length exam under real conditions.
Start free
